AW: AW: Possible PREEMPT_RT live-lock / priority-inversion between FUTEX_CMP_REQUEUE_PI and FUTEX_WAIT_REQUEUE_PI

[Moritz KLAMMLER (FERCHAU)]

>>>> we're running Linux 6.6 with PREEMPT_RT on a single-core armv7l machine
>>>
>>> v6.6.109+?
>>
>> Yes, 6.6.122 to be precise.
>
> Just to satisfy my curiosity, are you running my stable v6.6-rt kernels
> or rolling your own? (Not that it would affect this bug at all, since my
> tree has the same issue).

Hi Clark,

it's a BitBake build.  At the time when I started this investigation, it
took 6.6.122 from linux-stable plus rt67 and then some other patches as
well, none of which should affect the RT part, though (hopefully).
Meanwhile, we've moved on to linux-stable 6.6.129 + rt70 patch.  I can't
find a tag for 6.6.112 in linux-stable-rt (maybe for good reasons?) but
v6.6.129-rt70 does exist.  I don't know why the project has been set up
to patch linux-stable instead of using linux-stable-rt in the first
place, to be honest.  But I'll take this as an opportunity to suggest to
the team switching to linux-stable-rt as upstream.  Thanks for your
efforts maintaining it.

KR Moritz